State-of-the-art Open Source LLMs, Fine Tuning & Other Things // Luke Marsden // IRL #68 Bristol

Описание: MLOps Community IRL Meetup #68! The Bristol MLOps Community talked to Luke Marsden, Founder of MLOps Consulting sponsored by Cleo.

//Abstract
AI expert Luke Marsden explored the latest advancements in open-source large language models (LLMs) and machine learning algorithms. He highlighted the fine-tuning intricacies and diverse applications, from generating business ideas to understanding scientific papers. Marsden also delved into quantized Lora and Axolotl config for memory efficiency and model quantization. Addressing ethical concerns, he emphasized the need for human oversight and potential future competition between corporations and OpenAI in utilizing these models. The conversation wrapped up with insights into model operations and the exciting prospects of GPT-4 Vision and multimodal models.

// Bio
CEO https://helix.ml - Private GenAI Platform. Owner http://MLOps.consulting. Founder http://MLOps.community. Hacker & Entrepreneur.

// Jobs board
https://mlops.pallet.xyz/jobs

// Related links


---------- ✌️Connect With Us ✌️------------
Join our Slack community: https://go.mlops.community/slack
Follow us on Twitter: @mlopscommunity
Sign up for the next meetup: https://go.mlops.community/register
Catch all episodes, Feature Store, Machine Learning Monitoring, and Blogs: https://mlops.community/

Connect with Demetrios on LinkedIn:   / dpbrinkm  
Connect with Luke on LinkedIn:   / luke-marsden-71b3789  

Timestamps:
[00:00] Open source AI technology is outpacing proprietary
[03:04] Mistral Seven B model shows impressive performance
[06:19] Language model predicts user behavior, process improvement
[12:13] Reducing precision of numbers to save memory
[14:55] Gen AI exploding, Chat GPT changing everything
[19:41] LLM arena ranks models against each other
[21:47] Interacting with language models using instruction tuning
[25:14] Fine tune models for specific business language
[29:28] Chat GPT for text fine-tuning pipeline
[31:58] Research on data chunk indexing and embeddings
[37:18] Text chunked, model generates Q&A pairs, fine-tuned, useful for understanding complex papers
[39:06] AI can read images and do more